## Connecting Deep Reinforcement Learning based Obstacle Avoidance with Conventional Global Planners using Waypoint Generators

Deep Reinforcement Learning has emerged as an efficient dynamic obstacleavoidance method in highly dynamic environments . It has the potential toreplace overly conservative or inefficient navigation approaches . However, the integration of the method into existing navigation systems is still an open frontier due to the myopic nature of Deep Reinforcer Learning .…

## RNN Transducer Models For Spoken Language Understanding

We present a comprehensive study on building and adapting RNN transducer(RNN-T) models for spoken language understanding . These end-to-end (E2E) models are constructed in three practical settings: a case where verbatimtranscripts are available, a constrained case where the only availableannotations are SLU labels and their values .…

## Secure S Hell Introducing an SSH Deception Proxy Framework

Deceiving an attacker in the network security domain is a well established approach . This paper proposes a framework for placing decoy elements through an SSH proxy . This would allow to deploy decoys on-the-fly without the need for a modification of the protected system .…

## Deep Down the Rabbit Hole On References in Networks of Decoy Elements

Deception technology has proven to be a sound approach against threats toinformation systems . Decoy elements are causing distraction and uncertainty to an attacker . Deception is meant to be complementingfirewalls and intrusion detection systems . Particularly insider threats may bemitigated with deception methods, says author .…

## Half Duplex Attack An Effectual Attack Modelling in D2D Communication

The visualization of future generation Wireless Communication Network WCN is the presumption of onward innovations, the fulfillment of userdemands in the form of high data rates, energy efficiency, low latency, and long-range services . In comparison to previous technologies, these technologies exhibit flat architecture, the involvement of clouds in the network, centralized architecture incorporating small cells which creates vulnerablebreaches initiating menaces to the security of the network .…

## Unitary Subgroup Testing

We present a novel structural property of Clifford unitaries . Namely, that their (normalized) trace is bounded by $1/\sqrt{2$ in absolute value . We show a similar property for the $q$-aryCliffords . This allows us to analyze a simple single-query identity test under the Clifford promise and show that it has (at least) constant soundness .…

## Can Differential Privacy Practically Protect Collaborative Deep Learning Inference for the Internet of Things

Collaborative inference has emerged as an intriguing framework for applying deep learning to Internet of Things applications . For mitigation of privacy risks, differential privacy could be adopted in principle . However, practicality of differential privacy forcollaborative inference under various conditions remains unclear.…

## A Sketch Based Neural Model for Generating Commit Messages from Diffs

In this paper we apply neural machine translation (NMT) techniques to convert code diffs into commit messages . The results highlight that thisimprovement is relevant especially for Java source code files, by examining twodifferent datasets introduced in recent years for this task .…

## Characterization of Android malware based on opcode analysis

The Android operating system is the most spread mobile platform in the world . Attackers are producing an incredible number of malware applications for Android . Our aim is to detect Android’s malware in order to protect the user . Our method is tested both on a laboratory database and a set of real data.…

## Exploring Machine Speech Chain for Domain Adaptation and Few Shot Speaker Adaptation

Machine Speech Chain integrates both end-to-end (E2E) automatic speechrecognition (ASR) and text-to speech (TTS) into one circle for joint training . We apply few-shotspeaker adaptation for the E2E ASR by using a few utterances from targetspeakers in an unsupervised way, results in additional gains .…

## Learning Graph Structures with Transformer for Multivariate Time Series Anomaly Detection in IoT

Many real-world IoT systems comprising various internet-connected sensorydevices generate substantial amounts of multivariate time series data . Critical IoT infrastructures, such as smart power grids and water distribution networks, are often targets of cyber-attacks, making anomalydetection of high research value .…

## Characterization of Android malware based on subgraph isomorphism

The Android operating system is the most spread mobile platform in the world . Attackers are producing an incredible number of malware applications for Android . Our aim is to detect Android’s malware in order to protect the user . Our method is tested both on a laboratory database and a set of real data.…

## Auxiliary Tasks and Exploration Enable ObjectNav

ObjectGoal Navigation (ObjectNav) is an embodied task wherein agents are tonavigate to an object instance in an unseen environment . Agents achieve 24.5% success and 8.1% SPL, a 37% and 8% improvement over prior state-of-the-art, respectively, on the HabitatObjectNav Challenge .…

## AR Based Half Duplex Attack in Beyond 5G networks

With the evolution of WCN (Wireless communication networks), the absolutefulfillment of security occupies the fundamental concern . In view of security, we have identified another research direction based on the attenuation impact of rain in WCN . An approach is initiated by an eavesdropper in which a securecommunication environment is degraded by generating Artificial Rain (AR), which creates an abatement in the secrecy rate, and the cybersecurity getscompromised .…

## Statistically significant detection of semantic shifts using contextual word embeddings

Detecting lexical semantic shifts in smaller data sets is challenging due to a lack of statistical power . Non-contextual word embeddings mask the variability present in the data . We propose an approach to estimate semantic shifts by combining contextual word embeddeddings with permutation-based statistical tests .…

## Explainability based Backdoor Attacks Against Graph Neural Networks

Backdoor attacks on neural networks represent a serious threat to neural network models . A backdoor attack with trigger injecting position selected by GraphLIME reaches over $84%$ attack successrate with less than \$2.5% accuracy drop . Backdoor attack on node classification task reaches over .…

## The virtual element method for the coupled system of magneto hydrodynamics

Virtual Element Method allows us to construct noveldiscretizations for simulating realistic phenomenon in magneto-hydrodynamics . We show that this VEM approximation will yield divergence freediscrete magnetic fields, an important property in any simulation in MHD . We present a model for magneticreconnection in a mesh that includes a series of hanging nodes, which we use tocalibrate the resolution of the method .…

## WNARS WFST based Non autoregressive Streaming End to End Speech Recognition

Attention-based encoder-decoder (AED) end-to-end (E2E) models havedrawn more and more attention in the field of automatic speech recognition . Autoregressive beam search decoding makes it inefficient for high-concurrency applications . WNARS achieves a charactererror rate of 5.22% with 640ms latency, to the best of our knowledge, which is the state-of-the-art performance for online ASR .…

## Image based Virtual Fitting Room

Virtual fitting room is a challenging task yet useful feature for e-commerce platforms and fashion designers . Existing works can only detect very few types of fashion items . We firstly used Mask R-CNN to find the regions of different fashion items, and secondly used Neural Style Transfer to change the style of the selected fashion items.…

## Sound Probabilistic Inference via Guide Types

Probabilistic programming languages aim to describe and automate Bayesianmodeling and inference . Modern languages support programmable inference, which allows users to customize inference algorithms by incorporating guide programs . For Bayesian inference to be sound, guide programs must be compatible with model programs .…

## Extended Parallel Corpus for Amharic English Machine Translation

This paper describes the acquisition, preprocessing, segmentation, andalignment of an Amharic-English parallel corpus . It will be useful for machinetranslation of an under-resourced language . The corpus is larger thanpreviously compiled corpora; it is released for research purposes .…

## Machine Learning Based on Natural Language Processing to Detect Cardiac Failure in Clinical Narratives

The purpose of the study presented herein is to develop a machine learning algorithm based on natural language processing that automatically detectswhether a patient has a cardiac failure or a healthy condition . The results show that the combinationof TFIDF and MLPNN always outperformed other combinations with all overall performance .…

## Large Deviations and Information theory for Sub Critical SINR Randon Network Models

The article obtains large deviation asymptotic for sub-critical communicationnetworks modeled as signal-interference-noise-ratio networks . We define the empirical mark measure and the empirical link measure, as well asprove joint large deviation principles(LDPs) for the two empirical measures on two different scales, i.e.,…

## BSTC A Large Scale Chinese English Speech Translation Dataset

This paper presents BSTC (Baidu Speech Translation Corpus), a large-scale Chinese-English speech translation dataset . This dataset is constructed based on a collection of licensed videos of talks or lectures, including about 68hours of Mandarin data, their manual transcripts and translations into English .…

## Three Decades of Deception Techniques in Active Cyber Defense Retrospect and Outlook

Deception techniques have been widely seen as a game changer in cyberdefense . In this paper, we review representative techniques in honeypots,honeytokens, and moving target defense, spanning from the late 1980s to 2021 . Techniques from these three domains complement with each other andmay be leveraged to build a holistic deception based defense .…

## COVID 19 Named Entity Recognition for Vietnamese

The current COVID-19 pandemic has lead to the creation of many corpora thatfacilitate NLP research and downstream applications to help fight the pandemic . Most of these corpora are exclusively for English . In this paper, we present the first manually-annotated COVI-19 domain-specific dataset for Vietnamese.…

## FACESEC A Fine grained Robustness Evaluation Framework for Face Recognition Systems

We present FACESEC, a framework for fine-grained robustness evaluation offace recognition systems . We study five facerecognition systems in both closed-set and open-set settings . We find accurate knowledge of neural architecture is more important than knowledge of the training data in black-box attacks .…

## A Mixed method Study on Security and Privacy Practices in Danish Companies

Increased levels of digitalization in society expose companies to new security threats . Exogenous forces like new regulations,e.g., GDPR and the global COVID-19 pandemic, pose new challenges for companies . Study shows a misalignment between software developers and management when it comes to the implementation of security and privacy measures, i.i.…

## Direct PoseNet Absolute Pose Regression with Photometric Consistency

We present a relocalization pipeline, which combines an absolute poseregression (APR) network with a novel view synthesis based direct matching module . We modify the rotation representation from theclassical quaternion to SO(3) in pose regression, removing the need for balancing rotation and translation loss terms .…

## The Single Noun Prior for Image Clustering

Self-supervised clustering methods have achieved increasing accuracy in recent years but do not yet perform as well as supervised classification methods . We hypothesize that the performance gap is due to the difficulty of specifying, without supervision, which featurescorrespond to class differences that are semantic to humans .…

## TRiPOD Human Trajectory and Pose Dynamics Forecasting in the Wild

Joint forecasting of human trajectory and pose dynamics is a fundamentalbuilding block of various applications ranging from robotics and autonomousdriving to surveillance systems . Predicting body dynamics requires capturingsubtle information embedded in the humans’ interactions with each other and with objects present in the scene .…

## AlephBERT A Hebrew Large Pre Trained Language Model to Start off your Hebrew NLP Application With

AlephBERT is a large pre-trained language model for ModernHebrew . It is trained on larger vocabulary and a larger dataset than any PLM before . We present new state-of-the-art results on multiple Hebrew tasks and benchmarks . We make our model publicly available, providing a single point of entry for the development of Hebrew NLPapplications.…

## A single gradient step finds adversarial examples on random two layers neural networks

Daniely and Schacham recently showed that gradient descent finds adversarialexamples on random undercomplete two-layers ReLU neural networks . We extend their result to the overcomplete case, where the number of neurons is larger than thedimension . In fact we prove that asingle step of gradient descent suffices .…

## Enabling Cross Domain Communication How to Bridge the Gap between AI and HW Engineers

A key issue in system design is the lack of communication between hardware, software and domain expert . A HW/SWco-design process of (reconfigurable) neural accelerators, therefore, is animportant sub-problem towards a common co-design methodology . The ultimate challenge is to define the constraints for the design space exploration on the system level – a task which requires deep knowledge and understanding ofhardware architectures, mapping of workloads onto hardware and the application domain, e.g.…

## Algorithmic Obfuscation for LDPC Decoders

The main idea of logic locking is toinsert a key-controlled block into a circuit to make the circuit functionincorrectly without right keys . In the case that the algorithmimplemented by the circuit is naturally fault-tolerant or self-correcting, existing logic-locking schemes do not affect the system performance much evenif wrong keys are used .…

## Re designing cities with conditional adversarial networks

This paper introduces a conditional generative adversarial network to redesign a street-level image of urban scenes by generating an urban intervention policy, an attention map that localises where intervention is needed . The trained model shows strong performance in re-modellingcities, outperforming existing methods that apply image-to-image translation in other domains that is computed in a single GPU .…

## On Biasing Transformer Attention Towards Monotonicity

Many sequence-to-sequence tasks in natural language processing are roughlymonotonic in the alignment between source and target sequence . In this work, we introduce amonotonicity loss function that is compatible with standard attentionmechanisms . Performance is mixed, with larger gains on top of RNN baselines .…

## A Simple Geometric Method for Cross Lingual Linguistic Transformations with Pre trained Autoencoders

Powerful sentence encoders trained for multiple languages are on the rise . We investigate the use of a geometric mapping in embedding space to transformlinguistic properties without tuning of the pre-trained sentence encoderor decoder . We validate our approach on three linguistic properties using apre-trained multilingual autoencoder and analyze the results in bothmonolingual and cross-lingual settings .…

## Towards End to End Neural Face Authentication in the Wild Quantifying and Compensating for Directional Lighting Effects

The recent availability of low-power neural accelerator hardware, combined with improvements in end-to-end neural facial recognition algorithms provides,enabling technology for on-device facial authentication . The present researchwork examines the effects of directional lighting on a State-of-Art(SoA) neuralface recognizer . Top lighting and its variants (top-left, top-right) arefound to have minimal effect on accuracy .…

## Detection of Message Injection Attacks onto the CAN Bus using Similarity of Successive Messages Sequence Graphs

An attacker can inject messages (e.g.,increase the speed) that may impact the safety of the driver . This paper proposes a messageinjection attack detection mechanism independent of the IDs of the ECUs . The detection accuracy of the methods using a dataset collected from a moving vehicle undermalicious RPM and speed reading message injections show accuracy of98.45% when using LSTM-RNN .…

## A unified Abaqus implementation of the phase field fracture method using only a user material subroutine

We present a simple and robust implementation of the phase field fracture method in Abaqus . Unlike previous works, only a user material (UMAT) subroutine is used . This is achieved by exploiting the analogy between the Phase Fieldbalance equation and heat transfer, which avoids the need for a user elementmesh .…

## Learning What To Do by Simulating the Past

Recent work proposed that agents have access to a source of information that is effectively free: in any environment that humans have acted in, the state will already be optimized for human preferences . Such learning is possible in principle, but requires simulating all possible past trajectories that could have led to the observed state .…

## Probing BERT in Hyperbolic Spaces

This work considers a family of geometrically special spaces, the hyperbolic spaces, that exhibit better inductive biases for hierarchical structures . We argue that a key desideratum of a probe is its sensitivity to the existence of linguistic structures . In a syntactic subspace, our probe recovers tree structures than Euclidean probes, revealing the possibility that the geometry of BERT syntax may not necessarily be Euclidan .…

## A Survey on Security Issues of 5G NR Perspective of Artificial Dust and Artificial Rain

5G NR (New Radio) incorporates concepts of novel technologies such asspectrum sharing, D2D communication, UDN, and massive MIMO . This paper intends to provide an ample survey of security issues and their countermeasures encompassed in the technologies of 5GNR . Half-duplex attack specifies the attack solely on the downlink to spoof the allocated resources, with reducedmiss-rate.…

## Evolutionary rates of information gain and decay in fluctuating environments

In this paper, we wish to investigate the dynamics of information transfer inevolutionary dynamics . We use information theoretic tools to track how much information an evolving population has obtained and managed to retain about different environments that it is exposed to .…

## Displacement Driven Approach to Nonlocal Elasticity

This study presents a physically consistent displacement-driven reformulation of the concept of action-at-a-distance, which is at the foundation of nonlocalelasticity . The (total) strain energy is guaranteed to be convex and positive-definite without imposing any constrainton the symmetry of the kernels .…

## Learning to Coordinate via Multiple Graph Neural Networks

MGAN for collaborative multi-agentreinforcement learning is a new algorithm that combines graph convolutionalnetworks and value-decomposition methods . MGAN learns the representation of agents from different perspectives through multiple graph networks, and realizes the proper allocation of attention between all agents .…

## Weight Distributions of Two Classes of Linear Codes with Five or Six Weights

In this paper, based on the theory of defining sets, two classes offive-weight or six-weight linear codes over Fp are constructed . The weightdistributions of the linear codes are determined by means of Weil sums and anew type of exponential sums .…

## Advances in Metric Ramsey Theory and its Applications

Metric Ramsey theory is concerned with finding large well-structured subsetsof more complex metric spaces . For finite metric spaces this problem was first studied by Bourgain, Figiel and Milman . In this paper we provide deterministicconstructions for this problem via a novel notion of \emph{metric Ramseydecomposition}.…

## Bootstrapping of memetic from genetic evolution via inter agent selection pressures

We create an artificial system of agents (attention-based neural networks) selectively exchange messages with each-other in order to study memetic evolution and how memetic evolutionary pressures interact with genetic evolution of the network weights . In this system there is very little interaction between this memetic’ecology’ and underlying tasks driving individual fitness .…